local-browser

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Outsider free text can be ingested when the skill uses zero local-browser page snapshot / page selection to read the current page content from the user’s already paired browser, which may include arbitrary third-party web page text authored by outsiders.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill requires running npx -y -p @vm0/cli (which downloads and executes the @vm0/cli package from the npm registry at runtime), so it fetches and runs external code that the skill depends on.